SFM users have expressed the desire to transfer social media collections into preservation environments or repositories. To support this, collection portability has been implemented. SFM collection portability does not specify how the files must be transferred; WASAPI may be an approach for the file transfer.
Two aspects of collection portability don't quite map to the existing WASAPI spec:
A collection includes additional metadata files that are not WARC files.
The file structure of the collection must be maintained.
2. Export
SFM supports exporting derived social media datasets. Currently, users request exports using the UI. Export creation is asynchronous and users are notified by email when the exports are completed. Exports are supported in a variety of formats including Excel, CSV, and JSON. An export may be composed of multiple files and will always contain a README text file.
SFM export seems to map cleanly to the existing WASAPI spec. One possible issue is that content-type is not sufficiently expressive to indicate the role of a file. For example, it would not distinguish between the README text file and a dataset file.
The WASAPI spec aligns with two possible use cases for Social Feed Manager.
1. Collection transfer
SFM users have expressed the desire to transfer social media collections into preservation environments or repositories. To support this, collection portability has been implemented. SFM collection portability does not specify how the files must be transferred; WASAPI may be an approach for the file transfer.
Two aspects of collection portability don't quite map to the existing WASAPI spec:
2. Export
SFM supports exporting derived social media datasets. Currently, users request exports using the UI. Export creation is asynchronous and users are notified by email when the exports are completed. Exports are supported in a variety of formats including Excel, CSV, and JSON. An export may be composed of multiple files and will always contain a README text file.
SFM export seems to map cleanly to the existing WASAPI spec. One possible issue is that content-type is not sufficiently expressive to indicate the role of a file. For example, it would not distinguish between the README text file and a dataset file.